Ok, I ran into that just some minutes ago, so +1 for fix ;) On Tue, Aug 4, 2009 at 10:31 AM, Vidar Ramdal <vi...@idium.no> wrote:
> 2009/8/4 Carsten Ziegeler <cziege...@apache.org>: > > Bertrand Delacretaz wrote: > >> On Tue, Aug 4, 2009 at 10:08 AM, Carsten Ziegeler<cziege...@apache.org> > wrote: > >>> Bertrand Delacretaz wrote: > >>>> ...if a servlet is registered with > >>>> sling.servlet.extensions = html, it should not process any request > >>>> that doesn't have an html extension. > >>>> > >>> Hmm not sure :) I guess this boils down to the question if the > properties > >>> are applied using "or" or "and" logic. Atm it's "or" (at least this is > >>> my understanding of the code...) :) > >> > >> I don't see a use case for an OR between extensions and anything else, > >> to me it seems much more intuitive to use the extensions property to > >> limit the range of requests that a servlet handles. > >> > > Yes, maybe you're right - but nevertheless it's the way it has been > > for the past releases and people might rely on this behaviour. > > Looking at SLING-512 [1], where similar symptoms are described. > SLING-512 was fixed, but at some point ServletResourceProviderFactory > has been rewritten, and the bug was reintroduced. > So I guess I'd say fixing the current issue is in fact keeping > backwards compatibility - we're just going a bit further 'back' :) > > I'm attaching a proposed patch to SLING-1069 [2] - please have a look, > and preferably test it. As Carsten says, this has potential to break a > lot of stuff. > > [1] https://issues.apache.org/jira/browse/SLING-512 > [2] https://issues.apache.org/jira/browse/SLING-1069/ > > -- > Vidar S. Ramdal <vi...@idium.no> - http://www.idium.no > Akersgata 16, N-0158 Oslo, Norway > +47 21 531941, ext 2070 >